Impact of lack-of-benefit stopping rules on treatment effect estimates of two-arm multi-stage (TAMS) trials with time to event outcome

Table 8 Bootstrap results for trials that reached the final stage based on ICON4 and RE01

		I and D outcomes are OS^a				I outcome is PFS^b, D outcome is OS^a
		Estimated H R _Dfor trials reached final stage				Estimated H R _Dfor trials reached final stage
α ₁ ^c	δ ₁ ^d	e ₁ ^e	%Pass^f	Mean (centiles)	%Bias	e ₁ ^e	%Pass^f	Mean (centiles)	%Bias
a) ICON4 –		H R _D(95% bootstrap CI): 0.82(0.67,0.99)
0.50	1.00	75	99	0.82(0.67,0.98)	0	75	99	0.82(0.67,0.99)	0
0.40	0.96	98	97	0.82(0.67,0.97)	0	97	99	0.82(0.67,0.99)	0
0.30	0.94	126	96	0.82(0.67,0.97)	0	125	98	0.82(0.67,0.98)	0
0.20	0.91	163	92	0.81(0.67,0.94)	− 1	162	96	0.81(0.67,0.97)	1
0.10	0.89	222	86	0.81(0.67,0.94)	− 1	221	89	0.81(0.67,0.96)	1
b) RE01 –		H R _D(95% bootstrap CI): 0.75(0.60,0.93)
0.50	1.00	51	87	0.74(0.60,0.90)	−1	49	93	0.75(0.60,0.92)	0
0.40	0.96	66	89	0.74(0.60,0.90)	−1	64	92	0.75(0.60,0.92)	0
0.30	0.92	85	94	0.74(0.60,0.90)	−1	83	96	0.75(0.60,0.92)	0
0.20	0.89	110	99	0.71(0.59,0.80)	− 5	108	99	0.75(0.60,0.92)	0
0.10	0.86	150	89	0.73(0.60,0.86)	− 3	148	98	0.75(0.60,0.92)	0

^aOverall survival;
^bprogression-free survival;
^cone-sided significance level at stage 1;
^dpredefined threshold at stage 1;
^ecumulative number of control arm events required at end of stage 1;
^fpercentages of trials that pass the interim stage and continue accrual to the final stage.